Vivid Audio - Tech Talk

The Downside of the Wooden Square Box
Perhaps the most obvious aspect of Vivid Audio loudspeakers is that they come in enclosures that aren’t the typical square box but there’s a good reason for that and it’s not just an aesthetic decision. In a classic series of experiments back in the 1930s, Dr. Harry Olsen mounted a small full range driver in a number of wooden shapes, all having an approximately equal size. He clearly showed that the smoothest response was obtained when the enclosure was a large sphere. He showed equally clearly that one of the worst shapes is a rectangular prism. Despite this landmark work, most loudspeakers today are still made using sheet material cut and joined into rectangular boxes simply because of the cost. At Vivid Audio we made the decision, from the outset, that our products should not be compromised by such values and sought out manufacturing materials and techniques that would afford us complete freedom to design enclosures which give the best results. So our cabinets have very smooth curves around the tweeter and midrange unit.

Another downside to the rectangular box is the performance of the cabinet walls themselves. Any sheet of material will resonate at a particular set of frequencies and it is imperative that these fall outside the working range of the driver enclosed. Clearly the material should be as stiff and light as possible but much can be gained by the way in which it is used. One way of pushing the modal frequencies up is to curve the panel and yet another is to support the panel at regular intervals. Some designs of sheet material wooden boxes use a matrix of intersecting perforated wooden panels which span between the opposite faces. This works very well across the short dimension but between the top and bottom the webs are rather too long to be effective. Vivid enclosures are doubly curved everywhere and tapered towards both top and bottom so the area of these zones is small and the curvature tight thus obviating the need for top to bottom bracing. Between the sides however, bracing is of considerable benefit and is included in every gap between each driver and port. Adopting cast or vacuum-infused composite materials not only gives us freedom with the shape but also permits finishing in any of the huge range of sprayed paint finishes.

Acoustically Designed Cabinets – Mid and High Frequencies
In order to understand better why the Vivid Audio philosophy of using enclosures with rounded contours leads to an improved performance when compared to those with sharp corners we have to look at how sound behaves in the two cases. First let’s look at the way the rounded enclosure helps with the high frequencies. Sound energy is carried through air by means of pressure waves. The way in which these waves interact with their surroundings is not always intuitive and one of the most difficult ideas to grasp is why a wave radiating from a loudspeaker is reflected by a sharp cabinet edge. We’re all familiar enough with sound bouncing off a hard surface, that’s the echo you hear if you shout at a smooth wall. Imagine putting two flat boards either side of the tweeter; the sound bounces off the walls like ripples in a pond.

But there’s no hard surface at the edge of the cabinet, just free space, so what’s going on? As the sound moves outwards from the tweeter it forms hemispherical waves with one edge running along the flat cabinet front, sometimes referred to as half-space, until it reaches the cabinet edge and the full-space beyond. It turns out that the sudden change from being bounded on one side to being completely free is almost as big a shock as running into a hard surface and a sort of negative echo radiates from the edges

Re-radiation from sharp discontinuities, known as diffraction, is also something that concerns designers of sneaky military aircraft. They aim to create as little radar reflection as possible to avoid detection. The B2 bomber has a radar signature just one per cent of that of its less discreet predecessor the B-52. Much of this benefit has been achieved by keeping the surface as smooth as possible.

So how does this re-radiation affect the performance of the loudspeaker?

Some of the sound clearly takes the direct path to the listener’s ears and some of the sound bounces off the corners and heads back towards the listener’s ear where it mixes with that coming directly from the driver. Sometimes the waves add and sometimes they cancel depending on the wavelength and the position of the listener. This ‘interference’ causes the irregularities in the response shown by Olsen.

In the case of the curved cabinet there is no sharp edge and hence no interference.

Clever crossover design and other tricks can lessen the effects of interference directly in front of the speaker but when you place the speaker in a normal room with reflective walls you hear the main sound followed by a sort of average of all the irregular off-axis output so the subjective result is still coloured by the sharp corners.

Replacing the flat baffle and sharp edge with a single smooth curve means there is no longer a single point at which the sound space changes from half space to full space. In turn this means that there is no interference and a smooth off axis response.

Another frequently encountered shortcoming of many speaker systems is a jump in the dispersion between the different drivers. A typical cone mid-range driver concentrates the sound into an ever tighter beam as the frequency approaches the point at which it crosses over to the tweeter(2-3kHz). However a small tweeter mounted on a flat baffle, or worse still in free space, has a very wide dispersion at the lower end of its range. So if the response is flat through crossover when heard on-axis then there will be an excess of high frequencies off-axis. This excess reaches the listener after reflecting from the walls of the room and the subjective effect is of an over-bright balance.

Acoustically Designed Cabinets – Low Frequencies
At the bass end of the spectrum it’s no longer the wave-like nature of sound which is a concern but the sheer movement of the air. Vivid Audio loudspeakers use the vented style of enclosure to improve low frequency distortion performance. This means introducing a carefully optimised duct which connects the inside of the box to the outside. Sound output from the rear of the bass unit causes the air in this port to move in and out and for a certain band of low frequencies the majority of the system output is from this port. Benefits include reduced cone excursion and hence distortion but only if the air in the port moves smoothly. All too often vented loudspeaker systems have simple lengths of tube with no attention paid to the manner in which these are terminated. As the air moves back and forth through such tubes a distinct ‘chuffing’ sound may be heard that becomes rapidly more significant as the drive level is increased. The cause of this undesirable distortion is turbulence and it represents a transformation of the desired airflow into useless eddies. Turbulence occurs whenever there is an abrupt change in the direction of a flowing fluid. You can see it as you swish you hand through water or in the wake of a boat and you can hear it when you hiss through clenched teeth or in the roar of a jet engine. In every case the energy in the smooth stream is being converted into rapidly spinning vortices where it dissipates into waste heat. A quick look at any design where fluid movement is an issue reveals a simple solution – streamlining. Be it plane or a fish, sharp steps are nowhere to be seen. And so the same can be applied to loudspeaker ports. Vivid Audio loudspeakers feature bass ports with gentle flares at both the outside and inside ends and deliver a bass performance that far surpasses any simple tube.

CDP - Catenary Dome Profile
At Vivid Audio we believe in using pistonic drivers for every part of the spectrum because we have found, though experience and experiment, that no matter how well controlled it may be, cone breakup is always audible as a colouration of the sound. There may well be times when that colouration produces a pleasant effect but we feel that effects should be included by the artist in the recording and not be added later during the reproduction. To this end, all Vivid diaphragms have been optimised using computeraided finite element analysis to give the highest possible first break-up frequencies. In the case of the mid bass C125 cone diaphragm, this process yielded a solution with an unusually shaped central dome which is not so much a dust cap as an integral structural element. In the more critical dome diaphragms of the mid-range D50 and D26 HF driver, there were two phases of improvement to the standard metal dome. Using a ring of high modulus carbon fibre around the edge of an ordinary spherical aluminium dome was an idea hit upon twenty years ago by designer Laurence Dickie. The effect is to push up the first break-up frequency by an appreciable factor and this simple trick has since been exploited by a small number of manufacturers but Vivid Audio have taken the process to the next stage, re-optimised the profile to extract the maximum benefit from this technique. The solution takes the form of a rotated catenary, familiar as the natural form taken by a chain suspended at both ends, and yielded first break-up modes almost one octave higher than previously achieved using a simple spherical aluminium shell. Vivid Audio is unique in using the reinforced catenary dome for the mid-range frequencies and with a first break-up of over 20kHz the purity of reproduction of this critical band is guaranteed. The refined and patented technique offers an almost twofold improvement in performance over the ordinary metal dome. All diaphragms are made of anodised aluminium alloy which represents the best combination of stiffness and density when compared to titanium and magnesium, and we feel an optimum price/performance factor when compared to more exotic elements.

HAC - Highly Aligned Chassis
Fundamentally, the chassis of a loudspeaker maintains the relative position of the magnet, suspensions, voice coil and cone assembly. Ideally there should be nothing behind the cone which might cause resonance, reflection or any other perturbation of the sound coming from the rear of the diaphragm as this will affect its free motion and thus the sound coming from the front. Many early chassis took the form of a pressed metal dish with a few punched holes and, as a result, suffered badly from both resonance and reflection. Ironically, the most common early magnet systems used field coils or Alnico which tend to present a small area but this potential advantage was usually completely lost as a result of the simple chassis. Better engineered loudspeakers use die cast aluminium chassis but often with rather broad struts since this simplifies the design of the tooling. The advent of ceramic magnets was definitely a backward step from the point of view of rearward obstruction with magnets frequently having similar areas to that of the cone. Vivid Audio chassis are made from pressure die cast aluminium and have a unique construction in which the twelve supporting struts have an unusually narrow aspect ratio and are aligned in such a way as to present a minimum of obstruction. In fact, the 3mm wide legs of C125 have a total area of just 10% of the cone area which effectively renders them acoustically invisible. The compact rare earth radial magnet assembly fully complements this chassis with its small frontal area. A further function of the chassis is to act as a heatsink for the motor assembly and the cooler the motor runs, the smaller the power compression as the drive level is increased. By having a large number of chassis struts which are relatively deep next to their point of attachment to the magnet, the chassis is turned into a highly efficient finned heatsink almost tripling the total radiating area.

HVP - Highly vented Former
In a manner reminiscent of a piston in a cylinder, the central pole of the magnet in a standard moving coil driver alternately compresses and rarefies the air trapped within the voice coil, behind the dust dome. There is an escape route through the narrow gap between the coil and pole but the air experiences considerable aerodynamic drag which in turn damps the motion of the cone, particularly at low frequencies where the displacement is greatest. A common and simple way to solve this problem is to leave a large hole in the centre of the pole. The air now has an easy path to the outside world but it comes at a price for now there is a combination of an air volume and a connected duct. As with an open bottle or indeed a reflex loudspeaker system, the air in the duct reacts with that in the enclosed volume to form a Helmholtz resonator and at resonance the air in the tube reaches a peak in its motion while the movement of the cone is reduced. Typically the frequency of this resonance is around 300 to 400 Hz which is well within the bandwidth of the C125 bass/ mid driver. Adjusting the size of the hole within the limits of what is possible without affecting the magnet doesn’t improve things. Adding a row of hole around the coil former, however, was found to push up the resonance by a useful amount and to reduce its sharpness or 'Q' factor. Perforated formers are well known but have the disadvantage of causing audible noise as the air rushes through the small holes. If the number of holes is increased to the point where the hole area approaches half that of the total it turns out that the resonance is moved right out of band and the Q drops to the point where it becomes difficult to actually measure the resonance. Added to this is the fact that the air noise disappears because of the reduced velocity so the solution is near perfect.

Vivid Audio Filter Networks
All the crossovers in our loudspeakers are designed and built in-house to give the best results.

For a set of drivers to blend together seamlessly into a whole it is essential that their individual acoustic responses adhere accurately to certain defined filter shapes. We have chosen fourth order Linkwitz-Riley filters because of their ideal phase and summation characteristics through each crossover point which also yield symmetrical dispersion patterns. Thanks to the intrinsically smooth response of the drivers in their respective enclosures, the business of crossover design is made relatively straightforward but when combined with the use of computer aided analysis it has become possible to create complete sub-system responses which are astonishingly faithful to the ideal function; levels of accuracy which were once thought only possible with active systems. In the set of curves illustrated below, the response of the raw driver is shown together with a target response for the driver with filter. The result of adding a passive filter designed for a purely resistive load is quite far off the optimum showing a low output and peaky response. Running the circuit through computer optimiser dramatically improves the match to the target. The process is repeated for a second driver again resulting in a close match to the target response. When the two drivers and filters are run together the result is a near ideal summation. In common with both drivers and enclosures, all our crossovers are built in-house. This is becoming increasingly rare in this age of outsourcing everything but our reasons are completely pragmatic. In the early days of developing our first products, all the crossovers were built using inductors hand wound from oxygen-free copper mounted together with polypropylene capacitors on ply boards and hard wired together. Of course we assumed that once the designs were finalised we would use one of the well-established OEM suppliers as has become the industry norm. But while the production was impeccable, when we auditioned the first samples it was clear something was amiss. After rigorous checks that everything was exactly measurably identical it was clear the prototype gave a better subjective result and we decided that from that point all our crossovers would continue to be made by the same methods.

TTL - Tapered Tube Loading
In the early days of the quest for resonance and reflection free drivers it became abundantly clear that in order to fully complement the transparency of reproduction from a welldesigned dome transducer something had to be done to remove any hint of cabinet colouration. Colouration which could usually be traced back to resonant air pockets, the eigentones of the main cavity or structural modes in the enclosure walls. Mounting the drivers in the side of a large cylinder proved a very effective way of smoothing the path for the forward radiation but this did nothing for the sound coming out of the driver rear which has to be confined or dissipated in some way. Initial results, using a ring magnet within the voice coil diameter and a small central hole to connect the rear of the dome to the outside world, suffered from a strong Helmholtz resonance problem which imposed a severe peak and dip in the forward response. Replacing the magnet with an external ring and opening out the hole in the pole to the greatest diameter possible allowed the rear radiation to emerge unhindered but when attached to any sort of enclosure this performance was inevitably compromised by the resonant modes of the enclosed air space. Adding damping can improve this greatly but there had to be a better way. Coupling the driver instead to the end of a long tube excited a classic series of resonances but these were rather easier to damp out with a fibrous filling than with a short wide enclosure but it required rather a long tube to get good results. That wasn’t really an issue for a midrange unit but a 30cm bass driver in the end of a 3m tube was a rather different matter! So it was decided to try an exponentially tapered tube which, above its cut-off frequency, behaves pretty much like a straight tube but occupies about a third of the volume. And when adding the damping material it was found that simply drawing a length of fibre mat into the tube and letting the taper of the tube naturally compress the filling towards the narrow end resulted in a performance which actually exceeded that of the parallel tube. This principle has now been applied to both the high frequency and mid-range units of Vivid systems and ensures the best performance from both.

RCCM - Reaction Cancelling Compliant Mount
“Every action produces an equal and opposite reaction”. This simple Newtonion law of motion applies to all mechanical systems from rockets engines to loudspeaker motors. So when a current flows through the coil of a loudspeaker driver a force acts both upon it and the magnet surrounding it. The recoil velocity is proportionally smaller than that of the much lighter coil and cone and radiates a correspondingly small acoustic signal, that is, until it’s coupled to an enclosure having many times the area of the cone at which point its contribution can become significant. In addition to this simple area issue, if the cabinet suffers from structural resonances then the colouration may become a real problem. All of which could be avoided by decoupling the driver from the cabinet by the use of compliant mounts except that in order to be effective, the resonant frequency of the driver on its mounts must be well below the low frequency cut-off of the system. Holding the driver by its outer flange and achieving such a compliant mount are fairly mutually exclusive requirements and an alternative approach is necessary. A complete solution to the problem is to place identical drivers on opposite faces of the enclosure and couple the magnets rigidly together. So long as the drivers both receive an identical drive then the forces in the magnets will cancel completely and no motion will be transmitted to the enclosure. In our own reaction cancelling full range systems the two drivers are driven with identical signals below 100Hz. This means that above this frequency a differential force begins to appear so the driver pair must still be decoupled from the cabinet but the mount need not be as compliant as if it had to work down to 20Hz and is, in any case, easily implemented with the use of a pair of elastomeric o-rings under each driver rim. Our G1 and G2 Giya systems use identical drivers on the sides of the enclosure driver with an identical signal and complete reaction cancelling is assured.

RCP - Reaction Cancelling Ports
It is not usually appreciated that the air rushing in and out of the port in a vented loudspeaker also results in a small reaction force on the cabinet in the same way as a rocket moves in the opposite direction to that of the exhaust gases. A consequence of this is that the quality factor of the port resonance is affected by the mounting of the enclosure with respect to the ground. A good coupling, through spikes for example, or a complete decoupling on rubber mounts both result in a higher “Q” than when the enclosure is just resting on a carpet. All this vaguery can be eliminated by the simple act of using two opposing ports to cancel the cabinet reaction completely.

RTTL - Reflex Tapered Tube Loading
For some time now it has been recognised that fibre filled exponentially tapered tubes can deliver a performance which is very nearly that of an ideal enclosure. One which can contain the rear radiation from the drivers without any of the further resonant modes which plague ordinary loudspeaker cabinets. While being perfectly suited to mid-range and high frequency applications, where the rear of the driver diaphragm is coupled directly to the mouth of a horn with the same diameter, when it came to the low frequencies the tapered tube has certain shortcomings. Of course the approach can be used at low frequencies in a closed box mode but it is when we try to marry the advantages of the exponential absorber with those of reflex loading that the problems start. Reflex loading gives very real improvements in power handling and efficiency at the lowest frequency limit of the system which rely on the interaction between the air in the port, which connects the rear chamber to the outside world, reacting with the elasticity of the air contained in that chamber. If we take the example of a 320mm driver in a 200 litre box shown in the simulation below, the port output can clearly be seen summing with that of the driver and giving an ideal fourth order response. But further up the scale the horrors of the box resonances can also clearly be seen.

Adding fibre filling goes some way to improving the state of affairs at the high frequency end but also reduces the ‘Q’ factor of the port output and the whole system suffers. If we now take an exponentially tapered tube with a small amount of lossy filling which has the same diameter mouth as the diaphragm and the same total volume of 200 litres and attach it to the rear of the driver together with the same port as before, it is clear that the eigentones of the plain box have completely disappeared but so has the resonant port output. The system response has lost the ideal filter shape but also the driver excursion has increased. If we now increase the rate of taper of the horn, also known as the cut-off frequency, but retain the total volume, it is clear that the port output is improving but the resonances are still under control. And finally we find that if the horn cut-off is arranged to be four times that of the port tuning we completely restore the port output while the original cabinet resonances are still nowhere to be seen. It is this ideal enclosure design which is exploited in Giya to give a bass character which is so free of the clutter found in competitive systems. And it is the exponentially tapered horn which crowns Giya to such visually striking effect.

SFM - Super Flux Magnets
The use of radial magnets is still quite unusual in loudspeakers because of the difficulty in magnetizing in situ and the consequent need for novel ways of introducing live magnets into the steel assemblies. However the benefits are manifold. Vivid Audio leads the way in Tapered Tube loaded dome drivers and these rely on having the largest hole possible in the central pole to allow the sound to radiate freely into the absorber. So that excludes the use of a round disc magnet mounted within the voice coil. In addition to this there is always a benefit to be gained in terms of a smooth vertical dispersion by having the drivers close together. So that rather counts against having a ring magnet outside the coil. A rear mounted radial magnet neatly answers both requirements while offering further benefits. Because the poles of the magnet are polarized inside to outside rather that front and back, the design is selfshielding and can be used close to a traditional CRT monitor without further screening. Furthermore the naturally low leakage focussed field gives a higher maximum gap flux than a standard flat ring. A property which is exploited in the D26 tweeter which reaches a peak flux of 2.5T, roughly double that found on many 25mm tweeters, leading to a 96dB/W efficiency level.

Underhung Motor Systems
All the drivers used in Vivid Audio employ magnetic gaps which are longer than the voice coils so the coil is completely surrounded by steel throughout its operating range. This helps ensure that the coil inductance is stable irrespective of position but also, by being surrounded by thermally conductive metal, it helps keeps the coil temperature down which reduces the power compression. The penalty of this approach is that, particularly with the high excursion bass units, the driver needs more magnetic material but the benefits in terms of dynamics are so undisputable we believe it’s a price well worth paying.

System performance
By now you should have a fair impression of the way in which we at Vivid Audio have paid attention to every aspect of loudspeaker behaviour; the driver diaphragm, the coil, the magnet and the chassis. The cabinet material, the outside shape, the enclosure structure, the internal shape and driver loading. The crossover circuit, the components, the filter accuracy and the method of manufacture. Put together all these elements and you have one of the most surprising performances ever put on by a loudspeaker.